Inside the €500 Agent Audit, and what you leave with
For €500 you get a 90 minute call and a two page memo. Here is what happens in the 90 minutes, what the memo holds, and the terms.
Drafted by the same agent we build for clients. Reviewed and approved by a human.For €500 you get a 90 minute call and a two page memo. You walk away with a map of your workflows, a ranked shortlist of what an agent should do first, a named list of work to keep away from agents, and a memo a developer could start from.
The quick version:
- For €500 you get a 90 minute call and a written two page memo.
- The call maps your workflows, then ranks them by drudgery and by risk.
- You leave with a named list of work an agent should not touch.
- The memo names one agent to build first, its wiring, its main risk, and a build range.
- Money back if the memo is not useful, and the €500 comes off later work within 30 days.
Why the first conversation is priced
The first conversation is priced because a free audit still has to be paid for somewhere, and it is paid for by the project it exists to sell.
So the person running a free audit has one job, which is to come back with a reason you need the big build. You are unlikely to hear that you do not need this yet from someone whose income depends on the other answer.
Productized AI audits typically sell for €2,500 to €7,500, and most firms in the Netherlands and the wider EU give theirs away free. The Arrox offer ladder starts at €500, a fifth of the price at the bottom of that band.
That price does two things. It is small enough to approve without building a business case, and big enough that both sides prepare properly for the call.
We are paid for the diagnosis itself, so we earn no more by telling you to build bigger. If you are weighing an agent against a hire instead, what an agent costs against another hire covers the market rates behind that comparison.
The 90 minutes, beat by beat
The 90 minutes start from your real work, not from a discovery questionnaire. Beforehand you record a couple of Looms walking through the process as it runs today, and we watch those first, so none of the call goes on questions you have already answered.
That is the only preparation we ask for. The call itself runs in a set order.
Mapping the workflows you already run
Mapping the workflows takes the first stretch. We go through them one at a time: who starts it, what triggers it, which tools it touches, how often it runs, and where it sits waiting for a person.
The map is the first thing the call produces, and it is written down rather than described out loud. Seeing the chain in one place tends to move where you think the bottleneck is.
Ranking candidates by drudgery and risk
Ranking candidates uses two axes. Drudgery is how repetitive the work is and how often it runs. Risk is what happens if the agent gets it wrong and nobody notices for a week.
High drudgery with low risk is where you start. High risk with low drudgery is where you stay away, however good the demo looked. The rows below are illustrations, not a verdict on your own list.
| Drudgery | Risk if it goes wrong | Where it lands | Example |
|---|---|---|---|
| High, it repeats on a fixed schedule | Low, a person reads the output before it goes anywhere | Start here | A weekly report pulled together by hand |
| High, it repeats all day | Low, a human approves it before it ships | Start here | Drafts written for internal review |
| Low, it comes up occasionally | High, a wrong answer reaches a customer with nobody in between | Stay away | Answering customer questions unsupervised |
The ranking matters more than the map, because it decides where your first euro goes.
Naming what stays manual, and sketching the first build
Naming what stays manual comes before sketching any build. Three kinds of work come off the table: anything where a wrong answer reaches a customer with nobody in between, anything that is really about a relationship, and anything whose rules live in one person's head and have never been written down.
That last one is a writing job before it is an agent job. We say so on the call rather than quote you for it.
Sketching the build comes last. For the top candidate you get where it lives, usually Slack because that is where your team already is, what it may read, what it may write, where the human approval gate sits, and what the record of every run looks like afterwards.
The approval gate is the part that decides whether this is safe to put your name on. The Slack content agent we run in production stops at a human at every step, and the run does not continue until someone answers.
If most of your candidates are content jobs, what an agent can and cannot automate in a content team has the honest split. And if the answer is that nothing here is worth building yet, you hear it on the call rather than after a proposal arrives.
The two page memo, and what makes it build ready
The memo is two pages. A long deck is usually a way of hiding that the recommendation underneath it is thin.
Build ready means a developer, ours or yours, could pick it up and start work without another call. Four things are on the page:
- the single agent to build first, named
- the rough wiring, against the tools you actually use
- the main risk flag, usually wherever personal data enters or leaves
- a realistic build range, given as a range, because anyone quoting one number before scoping is guessing
The range is the part people skip past, and it is the honest part. A single figure quoted before anyone has seen your systems is a guess dressed up as a quote.
The memo is written to be forwarded, so you can send it to a co-founder, a CTO, or whoever holds the budget without translating it first.
The terms, plainly
The terms are short enough to state in two lines, and neither one needs a lawyer.
Money back if the memo is not useful. You decide that, not us, and there is no form to fill in.
The €500 comes off the price if you move up within 30 days. Continue and the diagnosis cost you nothing. Stop there and you spent €500 to find out where your team's hours go and which of them a machine has no business touching.
| Step | Price | What it is | Credit |
|---|---|---|---|
| Agent Audit | €500 | 90 minute call plus a two page memo | Comes off the next step within 30 days |
| Blueprint Sprint | €1,500 | Five days, a build ready plan | Credited against the build |
| Agent Build | €4,000 to €6,000 | Fixed scope, fixed price build | Whatever you paid for the Audit or Sprint comes off this number |
Whether the Audit is the right first step for you
The Audit is the right first step when you have a short list of jobs that might suit an agent and no reliable way to rank them.
It is the wrong first step when you already know exactly what to build and can describe it end to end. In that case the Blueprint Sprint is where to start, because turning a known build into a scoped plan and a fixed quote is what those five days are for.
If you are unsure which case you are in, the Audit is the cheaper mistake. The fee comes back if the memo is not useful, and comes off the next step if you continue.
FAQ
These are the questions founders ask before they book the first call.
What does the €500 Audit include?
The €500 Audit includes a 90 minute call and a two page memo written after it. The memo names the first agent to build, its rough wiring, its main risk flag, and a build range. You get the money back if you decide the memo is not useful.
Who needs to be on the call?
The call works best with whoever can describe the workflow end to end, and that is often one person. The preparation is a couple of Looms recorded before we speak, so nobody else has to sit in a session. Everyone else can read the memo afterwards.
What happens if the memo says we should not build anything?
The memo says exactly that when it is the honest answer, and you hear it on the call before it arrives in writing. You still keep the map of your workflows and the reasoning behind the answer. The money back terms apply either way.
Is the €500 credited if we go ahead with a build?
The €500 is credited against later work if you move up within 30 days. It comes off the €1,500 Blueprint Sprint or the €4,000 to €6,000 Agent Build. Continue and the diagnosis effectively cost you nothing.
How do I get the money back if the memo is not useful?
The money back is your call, and you get it by telling us the memo was not useful. There is no form to fill in and nobody on our side decides it for you. We would rather refund the fee than argue about it.
Can an agent be built without a Audit first?
An agent can be built without a Audit when you already know what to build and can describe it end to end. In that case the Blueprint Sprint is the better starting point, because it turns a known build into a scoped plan and a fixed quote. The Audit exists for the case where your shortlist is still a guess.
Booking the Audit, and what to bring
Booking the Audit takes one step. Pick a time on the Arrox offer page, then bring the workflow that annoys your team most and a couple of short Looms of it running.
Ninety minutes later you will know whether that workflow is the one to start with. If it is not, you will hear which one is instead, or that none of them is worth building yet.